Materials and methods Site description and field sampling

Sampling was performed at the south side of a barrier island (Spiekeroog) in a temperate salt marsh of the Wadden Sea of Germany (53°45'N, 43°19'E) in March 2015 during ebb (10 °Cair). Vegetation of the pioneer zone consisted mainly of Suaeda

Figure 2.1 A & B Location of the sampling area in north-western German barrier islands and in salt marshes along Spiekeroog; C "Pio" in top view; D Soil profile of "Pio"; E "Low" in top view;

F Profile of "Low"

maritima were dominant in the lower salt marsh zone. Undisturbed soil samples were taken from the upper layer (0-5 cm) of a Salic Fluvisol (WRB, 2014) from "Pio"

and "Low" using a soil corer (ø 7 cm) and transferred in a 250 ml "CombiSart"-Filterholder (Sartorius, Germany) in original vertical orientation of the soil column (Figure 2.1). No sieving was done to avoid artificially enhanced mineralisation (Datta et al., 2014).

Table 2.1 Basic characteristics of the Ai/Ah-horizons from "Pio"- and "Low"-soils

Soil Corg Nt C/N pH CaCO3 Sand Silt Clay PB (CFE)¹

[%] [%] [%]

63-2000 µm

2-63 µm <2 µm [mg C g^-1]

"Pio" 3.78 0.36 10.3 7.1 1.1 26 41.2 32.8 0.20

"Low" 5.47 0.49 11.2 7.1 1.2 22.8 41.6 35.5 0.37

1 PB (CFE): Prokaryotic biomass C derived from chloroform fumigated extraction

Experimental design

Figure 2.2 Scheme of the experimental equipment for simulation tides. Arrows show the direction of air flow (controlled by Pump 1); center: microcosm with soil core on filter holder (long dashed line below the soil core) and maximum water height in "Flood"- and

"Tide"-treatments (dotted line); left: tube with NaOH solution for CO2 trapping; bottom right: additional circle for "Tide"-treatments containing a secondary pump controlled by an automated timer, a switch to control air flow direction through "Tide"-circle and a salt water reservoir.

Pump 1

Soil

NaOH Reservoir

Pump 2 Switch

Timer

laboratory experiment, in which two soils were exposed to three inundation frequencies. In total, 48 samples from "Pio" and "Low" were preincubated at 20°C for 4 weeks until establishment of a stable CO2-efflux under following treatments: For simulation of "Ebb", samples were kept at initial soil moisture from the field. Soil in

"Flood"-treatments was constantly covered with 100 ml of artificial seawater containing NaCl (27.5 g l^-1), MgCl2*6H2O (2.44 g l^-1), MgSO4*7H2O (3.24 g l^-1) and CaCl2*2H2O (1.5 g l^-1) (Karius & Machunze, 2011). Same seawater was used to periodically flood the soils in the third treatment ("Tide") every 12 h for a period of 6 h. Treatments with constant levels of water, i.e. "Ebb" and "Flood", consist of a microcosm unit, one membrane pump (pump 1) and one flask with 10 ml NaOH (1 M) connected via PVC-tubes (Figure 2.2) (see also Kuzyakov & Siniakina, 2001). All connections were kept airtight unit had three upper outlets and one at the bottom that was supported by a filter holder framed by four layers of perforated polyethylene (0.5 mm). Pump 1 constantly pushed air from the bottom into the microcosm.

Evolved CO2 from the soil was carried along by the air stream through the upper outlet into the NaOH-solution for trapping. CO2-free air was sucked out of the headspace through the NaOH-vessel towards Pump 1 closing the circle.

Treatments with changing water levels (i.e. "Tide") had the same design explained above but were connected to a secondary pump (Pump 2), an automatic timer and a flask to store the artificial saltwater. Pump 2 was activated by the automatic timer every 6 hours daily for one minute to push into or to release saltwater from the reservoir to the microcosm by air pressure. To flood the microcosm, Pump 2 increased the air pressure in the sea water reservoir and thereby pushed the water into the upper inlet. Since Pump 1 constantly put pressure onto the lower inlet, no saltwater could reach Pump 1. Moreover, no water could flow back to the reservoir due to a stable pressure ensured by membrane Pump 2. Switching the three-way valve in front of Pump 2 resulted in a reversed circle: Pump 2 then sucked water into the sea water reservoir by simultaneously increasing air pressure inside the microcosm and decreasing in the reservoir. During the whole experiment temperature was kept constant between 21- 23 °C.

The experimental design was focused on mimicking natural conditions in the salt marsh. Strong dilution during tidal cycles as in-situ disadvantage could be avoided using the presented experimental approach. The dominating exchange of water in salt marshes is in vertical directions (Nuttle, 1988). This was considered in the design by a single inlet at the bottom of the microcosm for water exchange (Figure 2.2). We assured constant and comparable quality of the water by using artificial saltwater. In this regard, we excluded stochastically differences in nutrient- and salt-concentrations in natural seawater as factors affecting CO2-efflux and mineralisation.

For instance, nutrient ratios have been shown to impact on the balancing between immobilisation and mineralisation (Guenet et al., 2010). Moreover, changes in salinity were found to both positively and negatively correlate with decomposition in salt marsh ecosystems and were therefore excluded in the experimental design (Craft, 2007; Hemminga & Buth, 1991). With a constant air temperature during the experiment, we avoided temperature-mediated changes in SOM decomposition (Davidson & Janssens, 2006). Samples were set as intact cores into the microcosms to preserve the initial geochemical and prokaryotic stratification found in salt marsh soils (Froelich et al., 1979; Koretsky et al., 2005). Finally, no sieving was performed to avoid temporal respiration flush due to destruction of soil aggregates and penetration of O2 in deeper soil layers or release of former enclosed nutrients (Datta et al., 2014; Veen & Kuikman, 1990).

Labelling

After preincubation, two labelling stock solutions of 150 ml were prepared for both soils containing a label of 20 kBq of ¹⁴C-Glucose each and 400 mg ("Pio") and 815 mg ("Low") unlabelled D(+)-Glucose. As this corresponds to about double the amount of prokaryotic derived carbon (Cp) in "Pio" and "Low" soils, both soils received glucose approximately two times their prokaryotic C content. Each microcosm was labelled separately. Half of the microcosms (i.e. 4*"Pio-Ebb", 4* "Low-Ebb", 4* "Pio-Flood", 4*

"Low-Flood", 4* "Pio-Tide" and 4* "Low-Tide") received 10 ml of the corresponding stock solution distributed homogeneously over the soil. "Tide"-treatments were labelled during ebb-cycle with more than three hours remaining in the current tidal

served as control for calculation of PE.